Learning by message passing in networks of discrete synapses

被引:111
作者
Braunstein, A [1 ]
Zecchina, R [1 ]
机构
[1] Abdus Salaam Int Ctr Theoret Phys, I-34100 Trieste, Italy
关键词
D O I
10.1103/PhysRevLett.96.030201
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
We show that a message-passing process allows us to store in binary "material" synapses a number of random patterns which almost saturate the information theoretic bounds. We apply the learning algorithm to networks characterized by a wide range of different connection topologies and of size comparable with that of biological systems (e.g., n similar or equal to 10(5)-10(6)). The algorithm can be turned into an online-fault tolerant-learning protocol of potential interest in modeling aspects of synaptic plasticity and in building neuromorphic devices.
引用
收藏
页数:4
相关论文
共 40 条
[1]  
AMALDI E, 1991, ARTIFICIAL NEURAL NETWORKS, VOLS 1 AND 2, P55
[2]   CONSTRAINTS ON LEARNING IN DYNAMIC SYNAPSES [J].
AMIT, DJ ;
FUSI, S .
NETWORK-COMPUTATION IN NEURAL SYSTEMS, 1992, 3 (04) :443-464
[3]   LEARNING IN NEURAL NETWORKS WITH MATERIAL SYNAPSES [J].
AMIT, DJ ;
FUSI, S .
NEURAL COMPUTATION, 1994, 6 (05) :957-982
[4]  
BALDASSI C, IN PRESS
[5]   STORAGE CAPACITY OF A MULTILAYER NEURAL NETWORK WITH BINARY WEIGHTS [J].
BARKAI, E ;
KANTER, I .
EUROPHYSICS LETTERS, 1991, 14 (02) :107-112
[6]   BROKEN SYMMETRIES IN MULTILAYERED PERCEPTRONS [J].
BARKAI, E ;
HANSEL, D ;
SOMPOLINSKY, H .
PHYSICAL REVIEW A, 1992, 45 (06) :4146-4161
[7]   TRAINING A 3-NODE NEURAL NETWORK IS NP-COMPLETE [J].
BLUM, AL ;
RIVEST, RL .
NEURAL NETWORKS, 1992, 5 (01) :117-127
[8]   Polynomial iterative algorithms for coloring and analyzing random graphs [J].
Braunstein, A ;
Mulet, R ;
Pagnani, A ;
Weigt, M ;
Zecchina, R .
PHYSICAL REVIEW E, 2003, 68 (03) :15
[9]   Survey propagation:: An algorithm for satisfiability [J].
Braunstein, A ;
Mézard, M ;
Zecchina, R .
RANDOM STRUCTURES & ALGORITHMS, 2005, 27 (02) :201-226
[10]   Optimal information storage and the distribution of synaptic weights: Perceptron versus Purkinje cell [J].
Brunel, N ;
Hakim, V ;
Isope, P ;
Nadal, JP ;
Barbour, B .
NEURON, 2004, 43 (05) :745-757